Barker-ColloOF

JOURNAL et INTERPERSONAL
al. / COGNITIVE-BEHAVIORAL
VIOLENCE / AprilMODEL 2000
The goal of this study was to evaluate Joseph, Williams, and Yule’s cognitive-behavioral model
of response to traumatic stress when applied to a sample of 122 female sexual abuse survivors.
Participants completed surveys that measured each variable presented in Joseph, Williams, and
Yule’s model (i.e., event stimuli, personality, appraisals, coping, crisis support, event cognitions,
and emotional states). Path analysis showed that although Joseph, Williams, and Yule’s model
did not fit the data, a modified version based on the sexual abuse literature fit the data well.
Modifications to the model included the removal of the variable coping and the addition of paths
from event characteristics to crisis support and from personality to event characteristics.

A Cognitive-Behavioral Model
of Post-Traumatic Stress for
Sexually Abused Females
SUZANNE L. BARKER-COLLO
University of Auckland
WILLIAM T. MELNYK
Lakehead University
LESLIE MCDONALD-MISZCZAK
Western Washington University

As the study of sexual abuse has gained the interest of researchers, attempts
to explain individual variation in response to abuse have become increasingly
complex. A notable feature of recent research in the area of sexual abuse is the
increasing complexity of research designs and statistical techniques being
used. As noted by Alexander (1992), the future of the field is likely to depend
on “more complex models, hypotheses, and research designs” (p. 166). A
number of sexual-abuse-specific models have been proposed to explain indi-
vidual variation in response to sexual abuse using variables that are specific
to sexual abuse (e.g., Draucker, 1995; Wyatt, Newcomb, & Notgrass, 1991).
Unfortunately, the variables contained in these models do not lend them-
selves to modification through clinical interventions and are therefore

Authors’ Note: Sincere gratitude is expressed to each of the practitioners who assisted us in
obtaining participants for this study and to those women who gave their time in completing the
surveys. This research was supported by Doctoral Scholarship No. 752-97-1897 awarded by the
Social Sciences and Humanities Research Council of Canada.
JOURNAL OF INTERPERSONAL VIOLENCE, Vol. 15 No. 4, April 2000 375-392
© 2000 Sage Publications, Inc.
375
376 JOURNAL OF INTERPERSONAL VIOLENCE / April 2000

limited in clinical utility (e.g., frequency of sexual abuse incidents and maxi-
mum rapes per incident).
Whereas models specific to sexual abuse have been limited in utility,
Joseph, Williams, and Yule’s (1995) integrative cognitive-behavioral model
of response to traumatic stress (see Figure 1) might be used to successfully
examine individual variation in response to the trauma of sexual abuse. The
model includes variables that may be reliably measured and modified
through clinical intervention (e.g., coping and appraisals). As a model of
reactions to trauma that is not specific to a particular type of trauma, Joseph
et al.’s (1995) model allows reactions to sexual abuse to be placed within the
broader context of stress and coping. Components of the model include three
moderator variables (event stimuli, personality, and crisis support), two
mediators variables (event appraisals and coping), and two outcome or symp-
tom variables (event cognitions and emotional states).
Joseph et al.’s (1995) model of response to traumatic stress proposes a
complex pattern of interrelationships between model variables. The starting
point of this model is the experiencing of a traumatic event that presents the
individual with event stimuli. Event stimuli may be defined as the characteris-
tics of an event such as duration, frequency, and type of sexual abuse. Accord-
ing to the diagnostic criteria proposed by the American Psychiatric Associa-
tion (APA) (APA, 1994), salient characteristics of a traumatic event include
not only the overt characteristics of the event but also the degree of actual or
threatened death or serious injury and threat to physical integrity. Investiga-
tion of the relationship between specific aspects of sexual abuse and symp-
tom outcome indicate that sexual abuse that is more forceful and frequent,
more physically harmful or threatening, and that involves multiple perpetra-
tors results in greater symptomatology (Ellis, Atekeson, & Calhoun, 1981;
Kilpatrick et al., 1989; Meichenbaum, 1994; Neuman, Gallers, & Foy, 1989).
As described by Horowitz (1986), due to their salience, the stimuli pre-
sented in traumatic events must be processed in discrete portions. Processing
of event stimuli as event cognitions is generally thought to take the form of
intrusive thoughts, emotions, and behaviors (e.g., intrusive thoughts, flash-
backs, and nightmares) (Herman, 1992; Joseph et al., 1995). These event
cognitions parallel the intrusive phenomena identified as a symptom of post-
traumatic stress disorder (PTSD) (APA, 1994). There is evidence that these
intrusive phenomena are common and may represent a normal response to
traumatic events (Blank, 1993).
In relating event cognitions to other variables in the model, Joseph et al.
(1995) stated that these “traumatic cognitions, images, sounds, smells, and
tactile experiences will idiosyncratically reflect the individual’s prior experi-
ences” (p. 517). Event cognitions are therefore said to be moderated by
 Barker-Collo et al. / COGNITIVE-BEHAVIORAL MODEL 377

Figure 1: Joseph, Williams, and Yule’s (1995) Integrative Cognitive-Behavioral

Path Model of Post-Traumatic Stress Disorder.

personality variables. A variety of personality variables have been posited as

modifiers of intrusive symptomatology following sexual abuse. Although
Joseph et al.’s (1995) model includes the variable personality as a single con-
struct, one aspect of personality, neuroticism, is of particular interest in the
study of response to trauma due to its impact on response to stress. Neuroti-
cism has been defined as the predisposition to experience negative affective
states and behavioral manifestations of emotional instability (McCrae &
Costa, 1989). Bolger (1990) and Costa and McCrae (1992) identified neu-
roticism as an index of vulnerability to stressful events where persons who
score high on scales of neuroticism cope more poorly with stress than others.
Indeed, the presence of high levels of neuroticism has been linked to
increased emotional distress and reduced effectiveness of problem-solving
abilities following sexual abuse (Follette, Naugh, & Follette, 1997).
Event cognitions are also said to be influenced by appraisals. Appraisals
are defined by Joseph et al. (1995) as thoughts about the causation of trau-
matic events and the information depicted in event cognitions. Research indi-
cates that appraisals of events involving internal causal attributions (i.e.,
self-blame) are associated with increased depressive, anxious, and intrusive
symptomatology up to 2 years following trauma (Delahanty et al., 1997;
Joseph, Brewin, Yule, & Williams, 1991, 1993). In studies of attributions of
blame following sexual abuse, Wyatt et al. (1991) found that 65% (n = 55) of
female abuse survivors made internal attributions about the abuse event and
that this style of attribution was significantly related to negative symptom
378 JOURNAL OF INTERPERSONAL VIOLENCE / April 2000

outcomes. Joseph et al.’s (1995) model indicates that although appraisals

influence event cognitions, they are in turn influenced by personality.
In the model, the occurrence of event cognitions and appraisals that follow
exposure to event stimuli are proposed to elicit strong emotional states that
are themselves subject to further cognitive appraisals. Emotional states com-
monly identified as the long-term correlates of sexual abuse include depres-
sion, sadness, guilt, anxiety, and dissociation (Briere & Runtz, 1989; Finkel-
hor, 1990).
Event cognitions, appraisals, and emotional states are proposed to activate
attempts at coping. Although the mechanisms through which coping strate-
gies operate have not been clearly defined, few would deny their importance
in reducing anxiety and subjective distress following exposure to traumatic
events (Folkman & Lazarus, 1988). Investigators typically differentiate
between problem-focused coping and emotion-focused coping (Folkman &
Lazarus, 1988; Moos, 1993). A number of investigations have found that
greater reliance on emotion-focused coping strategies results in more severe
symptomatology following trauma (Cohen & Roth, 1987; Nezu & Carne-
vale, 1987; Z. Solomon, Mikulincer, & Flum, 1988). One important element
of coping identified by Joseph et al. (1995) is crisis support. Crisis support
has been defined by Joseph et al. (1995) as involving both the availability of
others and their reactions to disclosures of trauma. The evidence suggests
that, in general, individuals who receive higher levels of crisis support experi-
ence better psychological outcome (Joseph, Andrews, Williams, & Yule,
1992). Joseph et al.’s (1995) model indicates that crisis support is thought to
influence appraisals, coping, and emotional states.
Despite empirical support for and apparent clinical utility of variables
included in Joseph et al.’s (1995) model, although discrete portions of the
model have been empirically investigated (Joseph et al., 1991, 1993, 1996),
no attempt has been made to empirically evaluate the model as a whole. In
addition, those evaluations that have taken place have been limited to only
one sample (i.e., survivors of the Herald of Free Enterprise disaster).
The goal of this investigation was to evaluate Joseph et al.’s (1995) inte-
grative cognitive-behavioral model of response to trauma using data col-
lected from a sample of female sexual abuse survivors. For the purposes of
this study, sexual abuse has been defined as sexual contact ranging from pet-
ting and kissing to sexual intercourse and involving varying degrees of coer-
cion, threat, and force (Koss & Orso, 1982).
In testing Joseph et al.’s (1995) model, two modifications based on the lit-
erature regarding sexual abuse were hypothesized. The first modification
stems from Joseph et al.’s (1995) discussion of crisis support. In their discus-
sion, the authors noted general agreement in the literature (Jones & Barlow,
 Barker-Collo et al. / COGNITIVE-BEHAVIORAL MODEL 379

1990; Joseph et al., 1996; S. D. Solomon, 1986) that increased availability of

crisis support is predictive of external event appraisals (i.e., attribution of
blame to others) and reduced PTSD symptomatology. However, some trau-
matic events (e.g., sexual abuse) can be stigmatizing and elicit shunning,
avoidance, and blaming of the victim by crisis supports (Wortman & Leh-
man, 1985). Shunning, avoidance, and blaming by the support network and
failure to engage the network may be particularly evident where the abuser is
known to the survivor due to increased efforts to deny or hide the occurrence
of the abuse (Meichenbaum, 1994). Therefore, as the extent of crisis support
accessed by the survivor increases, so does the likelihood that some or all of
these crisis supports may react negatively. As such, it was hypothesized that
increased availability of crisis support will result in the initiation of negative
event appraisals (i.e., internal attribution of blame).
The second hypothesized modification arose in response to Joseph et al.’s
(1995) proposal that event characteristics will have a direct effect on only one
other variable, event cognitions. It has been found that some characteristics
of sexual abuse scenarios influence another variable within the model, the
engagement and provision of crisis support (Kilpatrick et al., 1989; Meichen-
baum, 1994; Parrot & Bechofer, 1991; Wyatt et al., 1991). Specifically,
engagement of crisis support has been linked to amount of force used by the
perpetrator. As indicated by Wyatt et al. (1991), increasing level of force used
by a perpetrator is significantly related to increasingly negative reactions of
others to the victim when sexual abuse is disclosed. Thus, the addition of a
path in the model from event stimuli to crisis support was hypothesized.

METHOD

Sample

Data were obtained from a sample of 122 female survivors of sexual abuse
from across the province of Ontario, Canada. Participants ranged in age from
15 to 57 years with a mean value of 31.4 years. Seventy-nine respondents
(64.8%) were Caucasian, whereas 39 (32.0%) were of Native American
ancestry. Of the remaining 5 participants, 3 were of Asian ancestry and 2
were of African ancestry. Education level ranged from grade 8 to completion
of a university degree. The majority of respondents (53.2%) had completed 1
year of university. The majority of respondents were single (41.3%), whereas
36 (29%) were married, 20 (16.4%) were divorced, 8 (6.6%) were in
380 JOURNAL OF INTERPERSONAL VIOLENCE / April 2000

common-law relationships, 6 (4.9%) were separated, and 2 (1.6%) were

widowed.
The abuse experienced by the respondents began at a mean age of 8.7
years. Respondents reported having experienced more than 50 separate inci-
dents of abuse in 30.3% of cases. Half of the sample were abused by an imme-
diate family member. The abuser was a member of the extended family or an
individual known to the victim in 49 (40.2%) cases. The remaining 12 indi-
viduals (9.8%) were abused by a stranger.

Measures

To ensure clarity, the names of the measure used in relation to each model
variable is noted in brackets below the name of the model variable in Figure 1.

Event stimuli. According to Joseph et al. (1995), event stimuli are the char-
acteristics of the traumatic event. As previously noted, sexual abuse that in-
volves greater force and greater frequency and is more physically harmful or
threatening is thought to result in more severe symptomatology (Ellis et al.,
1981; Kilpatrick et al., 1989; Meichenbaum, 1994; Neuman et al., 1989). In
the present study, the Sexual Experiences Survey (SES) (Koss & Orso, 1982)
was administered to assess the type and severity of sexual victimization. As
stated by Koss and Gidyaz (1985), the SES is “designed to reflect various de-
grees of sexual aggression and victimization” (p. 422) by assessing both the
type of sexual abuse (i.e., kissing, petting, intercourse, and oral or anal inter-
course) and the type of coercion or force used during abuse (i.e., threats of
physical force, verbal arguments, or physical force). The SES contains 13
items presented in a yes/no format (yes =1; no = 0) worded to portray female
victimization and male aggression. Scores for this scale are determined by
summing the total across responses, with a maximum score of 13. Koss and
Orso (1982) reported internal consistencies (Cronbach’s alpha) of .74
(women) and .89 (men) with a test-retest item agreement of 93%. For this
sample, mean performance on the SES was 8.22 with a standard deviation of
2.80. Cronbach’s alpha for SES items was .81. Although generally consid-
ered a measure of adult sexual victimization experiences, the format of the
SES, in which participants are asked, “Have you ever: . . . ?” followed by the
13 yes-no items, does not exclude its use for adults reporting sexual victimi-
zation experienced prior to the age of consent. Data reported on the develop-
ment, reliability, and validity of this measure make no reference to the age of
occurrence of the experiences reported (Koss & Gidycz, 1985; Koss & Oros,
1982).
 Barker-Collo et al. / COGNITIVE-BEHAVIORAL MODEL 381

For the purposes of the present study, 20 additional items were adminis-
tered separately from the SES to obtain additional descriptive information
from the sample. These items obtained information on relationship to the
abuser, age at which abuse occurred, number of episodes of abuse, and maxi-
mum rapes per incident. Scores obtained on these additional items were not
added to SES scores.

Personality. Although the model presented by Joseph et al. (1995) con-

tains overall personality functioning as a variable, the present investigation
examined a specific subtype of personality, neuroticism, due to its role in
regulating susceptibility to psychological distress and ability to cope with
stressful events (Costa & McCrae, 1992). Neuroticism was assessed using
the NEO Personality Inventory (NEO-PI) neuroticism scale (Costa &
McCrae, 1992), which contains six subscales of eight items each. Subscales
include anxiety, angry hostility, depression, self-consciousness, impulsive-
ness, and vulnerability. Each item is answered on a 5-point rating scale from
strongly disagree to strongly agree, with higher scores indicating greater lev-
els of neuroticism. Test-retest reliability for neuroticism is .87 (McCrae &
Costa, 1983). Mean performance on neuroticism for this population was
48.41 with a standard deviation of 10.58. Internal consistency for this scale
for the present sample was .74. Self-report measures on the neuroticism scale
are significantly correlated with peer ratings on the same factor (r = .54, p <
0.05) (McCrae & Costa, 1989).

Appraisals. According to Joseph et al. (1995), event appraisals are

thoughts about the information depicted when an individual reexperiences a
traumatic event (e.g., nightmares and flashbacks). Research indicates that ap-
praisals involving internal causal attributions are associated with more de-
pressive, anxious, and intrusive symptomatology up to 2 years following
trauma (Delahanty et al., 1997; Joseph et al., 1991, 1993). To determine
whether participants applied internal or external attributions, a modified ver-
sion of the Attributional Style Questionnaire (ASQ) (Peterson et al., 1982)
was administered. For the purposes of the present study, the wording of the
ASQ was modified from its original format to ask participants to respond in
reference to the most severe episode(s) of sexual abuse experienced (e.g.,
Was the sexual abuse episode(s) due to something about yourself or to some-
thing about the other people or circumstances involved?). Respondents rated
each of six items on a 7-point Likert-type scale (1 = totally due to others or
circumstances; 7 = totally due to me), with a maximum score of 42. All in-
structions and scoring were in accordance with Peterson et al. (1982). Inter-
382 JOURNAL OF INTERPERSONAL VIOLENCE / April 2000

nal consistencies for the ASQ are .75 for positive events and .71 for negative
events (Peterson et al., 1982). The mean level of response to this question-
naire for this sample was 34.02 with a standard deviation of 9.26. Internal
consistency for the scale in the present sample was .78.

Coping. Coping was measured using the Coping Responses Inventory

(CRI) adult form (Moos, 1993), a 48-item scale that measures eight different
coping types/scales of six items each. In responding to this measure, partici-
pants are typically asked to identify an event and complete the inventory in
reference to that event. For the purposes of the present study, participants
were asked to complete the inventory in reference to the most severe epi-
sode(s) of sexual abuse they experienced. Scales include logical analysis,
positive reappraisal, seeking guidance and support, problem solving, cogni-
tive avoidance, acceptance or resignation, seeking alternative rewards, and
emotional discharge. Each item is rated on a 4-point Likert-type scale that
ranges from 0 (no, not at all) to 3 (yes, fairly often) and that when summed
produces a maximum total score of 144. Scales are only minimally correlated
with social desirability (average absolute r = .13 for the eight scales) (Moos,
1993). Scoring procedures were in accordance with Moos (1993). Internal
consistency of the eight CRI scales for respondents ranged from .68 to .75.
Overall mean level of performance on this inventory was 53.91 with a stan-
dard deviation of 7.31.

Crisis support. Crisis support as defined by Joseph et al. (1995) was as-
sessed following the procedures of the Crisis Support Scale (CSS) of Joseph
et al. (1992). Using a 7-point Likert-type scale ranging from 1 (never) to 7
(always), the CSS assesses (a) availability of others, (b) contact with survi-
vors, (c) confiding in others, (d) emotional support, (e) practical support, (f)
negative response, and (g) satisfaction with support. Participants were asked
to respond to each item in relation to the most severe episode(s) of sexual
abuse they had experienced. The CSS has a maximum total score of 98, with
high scores indicating high levels of crisis support. Cronbach’s alpha for the
CSS for the current sample is .80. Mean level of performance on this measure
for this sample was 13.77 with a standard deviation of 8.77.

Event cognitions and emotional states. In the model, event cognitions are
defined as “re-experiencing phenomenon or intrusive recollections of the
trauma” (Joseph et al., 1995, p. 517), whereas emotional states refers to feel-
ings such as isolation/loneliness, anxiety, sadness, fear, inferiority, and guilt,
which often follow abuse. These two model variables were assessed using
those items of the Traumatic Symptom Checklist-40 (TSC-40) (Elliott & Bri-
 Barker-Collo et al. / COGNITIVE-BEHAVIORAL MODEL 383

ere, 1991) that fit with these definitions. Each of the TSC-40’s 40 items asks
the respondent to rate the frequency (never = 0; often = 3) with which she has
experienced a specific symptom. Participants were asked to rate each item in
relation to the sexual abuse they had experienced. The Cronbach’s alpha for
this sample was .90. Internal consistencies for event cognitions and emo-
tional states for the present sample were .78 and .71, respectively.

Procedure

Thirty-two agencies providing counseling services to sexual abuse survi-

vors from across Ontario, Canada, agreed to distribute survey packages
among their clientele. Due to the possible psychological effects associated
with participation in this study, practitioners were asked to agree to provide
psychological debriefing to clients if required during or following comple-
tion of the survey package. Data was gathered over a period of 6 months (May
1996 to December 1996). Participant survey packages included a cover letter
explaining the purpose, requirements, and the voluntary and confidential
nature of participation; all instructions and materials relevant to the instru-
ments used; and a stamped, self-addressed envelope. Agencies were asked to
invite all consecutive new referrals to participate. A total of 146 of the 417
distributed survey packages (35.8%) were returned. This response rate is
comparable to that obtained in similar examinations of response to sexual
abuse that used a survey method (Coffey, Leitenberg, Henning, Turner, &
Bennett, 1996). All respondents met the authors’definition of sexual abuse as
indicated by their responses to survey items. Of the returned surveys, 9 were
incomplete, 4 were completed by males, 2 were blank, and 2 were not legible.
The remaining 126 surveys were examined for accuracy of data entry, miss-
ing values, outliers, and the assumptions of multivariate analysis. Univariate
outliers are cases with extreme standardized scores (i.e., z scores) on one or
more variables. Cases with standardized scores in excess of ±3.00 are consid-
ered outliers. Three of the cases in this sample had extreme standardized
scores on one or more variables. These cases were excluded from further
analysis. The potential for multivariate outliers, or cases with an unusual pat-
tern of scores, was also examined. To identify multivariate outliers, mahala-
nobis distances are computed. Mahalanobis distance is the measure of the
difference between a single case and the central value for all other cases. If a
case has an unusual pattern of relationships between variables, the mahalano-
bis distance between that case and the central value for the remaining cases
will be significant. As the mahalanobis distance for one case within this sam-
ple was significant, this case was not included in any further analyses. Once
outliers were excluded, 122 cases remained in the data set.
384 JOURNAL OF INTERPERSONAL VIOLENCE / April 2000

RESULTS

Path analysis (LISREL 7.0) (Joreskog & Sorbom, 1988) was used to
examine Joseph et al.’s (1995) integrative cognitive-behavioral model of
PTSD (see Figure 1). As the present study examines cross-sectional data, it
was not appropriate to examine the bidirectional effects presented in the
model, which represent changes in relationships between variables that are
proposed to develop over time. Examination of relationships over time would
require the collection of longitudinal data. Due to the cross-sectional nature
of the data, all bidirectional paths within Joseph et al.’s (1995) model were
replaced with unidirectional paths to represent the first stage in the longitudi-
nal chain of events presented by Joseph et al. (1995). When tested, it was
found that this model did not fit the data, χ (9) = 27.8, p < .001; Goodness of
2

Fit Index (GFI) = 0.705.

When a model does not fit the data, modification indices generated by
LISREL 7.0 (Joreskog & Sorbom, 1988) can be used to guide modifications
to the model. Modification indices show the change in χ expected if a single
2

parameter/path was freed (if currently constrained) or constrained (if cur-

rently free). Using Raykov’s (1994) criteria, modification indices that
exceeded a value of 5 were considered. Two paths had modification indices
greater than 5. One of these indices was associated with addition of a path
from event characteristics to crisis support. The addition of this path was
hypothesized a priori on the basis of the sexual abuse literature (Wortman &
Lehman, 1985). The second modification index was associated with the addi-
tion of a path from personality to event characteristics. After freeing these
paths, the model fit the data, χ (7) = 11.27, p = .127. Addition of the two paths
2

did not endanger the interpretability of the model. The second hypothesized
modification to the model was a change in the sign of path from crisis support
to appraisals. In presenting their model, Joseph et al. (1995) agreed with the
general literature on stress and coping that increased crisis support is predic-
tive of external appraisals of blame. However, the sexual abuse literature indi-
cates that increased crisis support may elicit shunning, stigmatization, and
blaming of the victim. This second hypothesis was also supported.
Following the addition of the two paths identified by modification indices,
a number of paths with nonsignificant beta weights remained in the model.
To streamline the model, these nonsignificant paths were dropped. The
resulting streamlined model fit the data well, χ (13) = 13.41, p > .4; GFI =
2

.970, (see Figure 2). Regression analysis indicates that this model accounts
for 61.3% and 28.5% of the variance in the symptom variables emotional
states and event cognitions, respectively. The amount of unique variance
 Barker-Collo et al. / COGNITIVE-BEHAVIORAL MODEL 385

Figure 2: Modified Version of Joseph, Williams, and Yule’s (1995) Model Follow-
ing Alterations Based on Modification Indices and Removal of Paths With Non-
significant Beta Weights

associated with each path in the model is indicated by an R2 value in brackets

below the beta weight of that path.

DISCUSSION

The primary goal of this study was to evaluate Joseph et al.’s (1995) model
of PTSD when applied to a sample of 122 sexual abuse survivors and to
evaluate two hypothesized modifications to the model. In testing Joseph
et al.’s (1995) integrative cognitive-behavioral model, it was found that a
modified version of the model fit the data well. In the modified model,
force/extent of sexual abuse (event stimuli) was linked to greater fre-
quency/variety of event cognitions. In the sexual abuse literature, increased
force/extent of abuse has indeed been associated with poorer symptom
outcomes (Meichenbaum, 1994; Wyatt et al., 1991). The model also offers
evidence that in addition to characteristics of the event, factors such as per-
sonality, crisis support, and appraisals of blame are important in determining
individual variations in symptom presentation following sexual abuse.
Personality variables have also been hypothesized to mediate the relation-
ship between the experience of abuse and symptomatology. For example,
Joseph et al. (1995) stated that “intrusive ideation is . . . influenced by person-
ality and/or representations of earlier experience” (p. 517). Joseph et al.
386 JOURNAL OF INTERPERSONAL VIOLENCE / April 2000

(1995) also asserted that appraisals are thoughts about event cognitions that
draw on personality. Specifically, it has been proposed that women with low
self-esteem, low self-efficacy, or rigid role socialization may be more likely
to blame themselves for the occurrence of abuse (Bandura, 1986; Walker,
1984). Both direct and indirect paths in the model replicate this proposed
relationship. In the modified model, increased crisis support was linked to
internal appraisals of blame, which was in turn linked to reduced negative
emotional states. This finding is also supported in the sexual abuse literature.
In the sexual abuse literature, it has been found that some characteristics of
sexual abuse scenarios influence the engagement and provision of crisis sup-
port (Kilpatrick et al., 1989; Meichenbaum, 1994; Parrot & Bechofer, 1991;
Wyatt et al., 1991). Specifically, engagement of crisis support has been linked
to amount of force used by the perpetrator. As indicated by Wyatt et al.
(1991), increased levels of force used by a perpetrator was significantly
related to increased likelihood of negative reactions of others to disclosure of
the abuse. The findings of this study support this proposed relationship
between event stimuli to engagement of crisis support.
As the modified model fits with the majority of relationships proposed by
Joseph et al.’s (1995) integrative cognitive-behavioral model and with the
general literature on response to trauma, it appears that models of response to
trauma that place reactions to sexual abuse within the broader context of
stress and coping are applicable to sexually abused populations. However,
the findings also indicate that examination of sexual abuse within the larger
trauma response literature must take into consideration the uniqueness of
sexual abuse and its effects as a traumatic event. For example, although the
literature on response to trauma generally agrees that greater availability of
crisis support is predictive of positive event appraisals (i.e., external attribu-
tion of blame) and reduced symptomatology (Jones & Barlow, 1990; Joseph
et al., 1996; S. D. Solomon, 1986), it has been proposed that increased crisis
support following sexual abuse may have the opposite effect through shun-
ning, avoidance, and blaming of the victim by members of the crisis support
network (Wortman & Lehma, 1985). The findings of this study support the
presence of this unique relationship.
In understanding the relationships presented in the model, the literature on
PTSD must also be examined. For example, Blank (1993) and Horowitz
(1980, 1986) concurred that event cognitions are a normal response to trauma
that allow processing of traumatic information and lessen other negative
symptomatology. In the model, increased event cognitions were associated
with decreased negative emotional states. This suggests that processing trau-
matic information at a high rate results in a lessening of other negative symp-
tomatology, in this case, negative emotional states. Although some studies
 Barker-Collo et al. / COGNITIVE-BEHAVIORAL MODEL 387

(Wyatt et al., 1991) associate negative emotional outcomes with self-

blame, the results of the present study support the findings of Tennen and
Affleck (1990) who suggested blaming others results in poorer emotional
adjustment.
In the model, increased event cognitions were associated with reduced
negative emotional states. This raises the following question: Reduction of
which aspect of symptomatology should be the focus of clinical interven-
tion? According to Joseph et al. (1995), event cognitions are iconic represen-
tations of event stimuli. Due to their overwhelming nature, these representa-
tions are held in active memory for further conscious processing. As
described by Horowitz (1986), traumatic events must be processed in small
and discrete portions that allow the individual to maintain equilibrium.
The processing of event stimuli as event cognitions is generally thought to
take the form of intrusive thoughts and behaviors (e.g., dreams and flash-
backs) (Herman, 1992; Joseph et al., 1995). High levels of event cognitions
are thought to indicate that information about the traumatic event is being
processed at a high rate. Negative emotional states, on the other hand, have
been linked to an inability to process and cope with trauma. As increased lev-
els of event cognitions were indeed found to be related to reduced levels of
negative emotional states, the role played by event cognitions in the process-
ing of traumatic information would seem to be upheld. This interpretation of
the findings points toward the necessary role of event cognitions in coping
with trauma. Despite their apparent importance to processing of traumatic
information, event cognitions such as flashbacks and nightmares are
extremely disturbing symptoms. Perhaps in continuing to use techniques to
reduce these symptoms, clinicians should temper their use with an under-
standing of the role event cognitions appear to play in the processing of
event-related stimuli.

Limitations

The main limitation of this study is its reliance on self-report measures.

The use of self-report measures may result in bias due to the limiting format
in which questions are answered. Although some reduction in potential bias
effects were sought through ensuring confidentiality and anonymity of
responses, the findings would be strengthened by using additional measures
in future research. Specifically, reports by involved clinicians would provide
a basis for cross-validating the responses.
One related limitation is important to emphasize. The symptoms reflected
in the model are not unique to individuals with a history of sexual abuse. It
must therefore be emphasized that one cannot infer that someone who
388 JOURNAL OF INTERPERSONAL VIOLENCE / April 2000

presents with these symptoms has been sexually abused. The women in this
sample self-identified their history of sexual abuse before entering the study.
This study provides information on various clinical presentations and con-
cerns that may manifest themselves in this population.
The use of a clinical sample in conducting this study poses a number of
limitations to the generalizability of the findings. First, all of the sexually
abused women in this sample were given the opportunity to complete the sur-
vey through their contact with an organization or individual who provides
mental health services to survivors of sexual abuse. Because the sample was
obtained through respondents’ contact with mental health professionals, the
findings cannot be generalized to those women who have been sexually
abused but who have not come into contact with mental health services. Simi-
larly, whereas all consecutive referrals to each participating agency were
given the opportunity to participate in the study, only 35.8% of distributed
surveys were completed. The generalizability of the results is therefore lim-
ited only to those willing to complete the survey and cannot be extended to
those who did not wish to participate.
In addition, due to the need to ensure that completion of the survey did not
impact negatively on participants, clinicians involved in this study were
asked to provide participants with debriefings as required should participants
experience negative psychological effects as a result of their participation.
Although there is no data available to determine whether any of the respon-
dents required debriefing as a result of their participation, the availability of
debriefing services may have impacted on willingness to complete the
survey.
Finally, whereas this study tested a model of response to trauma, the data
collected were retrospective and cross-sectional, precluding the investigation
of the bidirectional relationships proposed. Future research should examine
the fit of the model or individual paths within the model when applied to
changes in response to sexual abuse over time. In addition, the size of the
sample was not adequate to allow cross-validation of the findings. Future
research should reevaluate the findings of this study.

Conclusion

The challenges that sexual abuse presents to society in general and to men-
tal health professionals in particular are enormous and complex. This study
provides a better understanding of the interrelationships between psychoso-
cial factors and symptom presentation in female survivors of sexual abuse.
The modified model presented here provides a valid framework for
 Barker-Collo et al. / COGNITIVE-BEHAVIORAL MODEL 389

understanding the impact of various factors on response to sexual abuse and

for planning interventions with survivors of sexual abuse. As a model of
responses to traumatic stress that is not specific to a particular type of trau-
matic event, the model also provides a framework for comparison of survi-
vors of sexual abuse and survivors of other forms of trauma. If replicated, the
results of this study could provide a context for understanding individual dif-
ferences and commonalties in response patterns and provide direction to cli-
nicians in how to approach these differences to better relieve negative symp-
tomatology. The need for further examination to validate and expand on the
findings of this study is warranted.

REFERENCES

Alexander, P. C. (1992). Introduction to the special section of adult survivors of childhood sexual
abuse. Journal of Consulting and Clinical Psychology, 60, 165-166.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disor-
ders (4th ed.). Washington, DC: Author.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Engle-
wood Cliffs, NJ: Prentice Hall.
Blank, A. S. (1993). The longitudinal course of posttraumatic stress disorder. In J.R.T. Davidson &
E. B. Foa (Eds.), Posttraumatic stress disorder: DSM-IV and beyond (pp. 3-22). Washing-
ton, DC: American Psychiatric Press, Inc.
Bolger, N. (1990). Coping as a personality process: A prospective study. Journal of Personality
and Social Psychology, 59, 525-537.
Briere, J., & Runtz, M. (1989). The Trauma Symptom Checklist (TSC-33): Early data on a new
scale. Journal of Interpersonal Violence, 4, 151-163.
Coffey, P., Leitenberg, H., Henning, K., Turner, T., & Bennett, R. T. (1996). The relationship
between methods of coping during adulthood with history of childhood sexual abuse and
current psychological adjustment. Journal of Consulting and Clinical Psychology, 64,
1090-1093.
Cohen, L. J., & Roth, S. (1987). The psychological aftermath of rape: Long-term effects and
individual differences in recovery. Journal of Social and Clinical Psychology, 5, 525-534.
Costa, P. T., & McCrae, R. R. (1992, August). What lies beneath the Big Five? Facet scales for
agreeableness and conscientiousness. In O. P. John (Chair), The Big Five: Historical per-
spective and current research. Symposium conducted at the meeting of the Society for Multi-
variate Experimental Psychology, Honolulu, HI.
Delahanty, D. L., Herberman, H. B., Craig, K. J., Hayward, M. C., Fullerton, C. S., Ursano, R. T., &
Baum, A. (1997). Acute and chronic distress and posttraumatic stress disorder as a function
of responsibility for survivors of motor vehicle accidents. Journal of Consulting and Clinical
Psychology, 65, 560-567.
Draucker, C. (1995). A coping model for adult survivors of childhood sexual abuse. Journal of
Interpersonal Violence, 10, 159-175.
390 JOURNAL OF INTERPERSONAL VIOLENCE / April 2000

Elliott, D. M., & Briere, J. (1991). Studying the long-term effects of sexual abuse: The Trauma
Symptom Checklist (TSC) Scales. In A. Wolbert Burgess (Ed.), Rape and sexual assault.
New York: Garland.
Ellis, E. M., Atekeson, B. M., & Calhoun, K. S. (1981). An assessment of long-term reaction to
rape. Journal of Abnormal Psychology, 90, 263-266.
Finkelhor, D. (1990). Early and long-term effects of childhood sexual abuse: An update. Profes-
sional Psychology: Research and Practice, 5, 325-330.
Folkman, S., & Lazarus, R. S. (1988). Coping as a mediator of emotion. Journal of Personality
and Social Psychology, 54, 466-475.
Follette, W., Naughe, A., & Follette, V. (1997). MMPI-2 profiles of adult women with childhood
sexual abuse histories: Cluster-analytic findings. Journal of Consulting and Clinical Psy-
chology, 65, 858-866.
Herman, J. L. (1992). Complex PTSD: A syndrome in survivors of prolonged and repeated
trauma. Journal of Traumatic Stress, 5, 377-392.
Horowitz, M. J. (1980). Psychological response to serious life events. In V. Hamilton & D. Warburton
(Eds.), Human stress and cognition: An information processing approach (pp. 235-263).
New York: John Wiley.
Horowitz, M. J. (1986). Stress response syndromes (2nd ed.) Northvale, NJ: Jason Aronson.
Jones, J. C., & Barlow, D. H. (1990). The etiology of posttraumatic stress disorder. Clinical Psy-
chology Review, 10, 299-328.
Joreskog, K. G., & Sorbom, J. (1988). LISREL 7: A guide to the program and its applications.
Chicago: SPSS Inc.
Joseph, S., Andrews, B., Williams, R., & Yule, W. (1992). Crisis support and psychiatric symp-
tomatology in adult survivors of the Jupiter cruise ship disaster. British Journal of Clinical
Psychology, 31, 63-73.
Joseph, S., Brewin, C., Yule, W., & Williams, R. (1991). Causal attributions and psychiatric
symptoms in survivors of the Herald of Free Enterprise disaster. British Journal of Psychia-
try, 159, 542-546.
Joseph, S., Brewin, C., Yule, W., & Williams, R. (1993). Causal attributions and psychiatric
symptoms in adolescent survivors of disaster. Journal of Child Psychology and Psychiatry,
34, 247-253.
Joseph, S., Dalgeleish, T., Thrasher, S., Yule, W., Williams, R., & Hodgkinson, P. (1996).
Chronic emotional processing in survivors of the Herald of Free Enterprise disaster: The
relationship of intrusions and avoidance at 3 years to distress at 5 years. Behavioral Research
and Therapy, 33, 1-4.
Joseph, S., Williams, R., & Yule, W. (1995). Psychosocial perspectives on post-traumatic stress.
Clinical Psychology Review, 15, 515-544.
Kilpatrick, D. G., Saunders, B. E., Amick-McMullen, A., Best, C. L., Veronen, L. J., & Resick, P. A.
(1989). Victim and crime factors associated with the development of crime-related posttrau-
matic stress disorder. Behavior Therapy, 20, 199-214.
Koss, M., & Gidyaz, C. (1985). Sexual Experiences Survey: Reliability and validity. Journal of
Consulting and Clinical Psychology, 53, 422-423.
Koss, M., & Orso, C. (1982). The Sexual Experiences Survey: A research instrument investigat-
ing sexual aggression and victimization. Journal of Consulting and Clinical Psychology, 50,
455-457.
McCrae, R. R., & Costa, P. T. (1983). Joint factors in self-reports and ratings: Neuroticism,
extraversion, and openness to experience. Personality and Individual Differences, 4, 245-255.
 Barker-Collo et al. / COGNITIVE-BEHAVIORAL MODEL 391

McCrae, R. R., & Costa, P. T. (1989). Reinterpreting the Meyers-Briggs type indicator from the
perspective of the five-factor model of personality. Journal of Personality, 57, 17-40.
Meichenbaum, D. (1994). A clinical handbook/practical therapist manual for assessing and
treating adults with post-traumatic stress disorder (PTSD). Waterloo, Ontario, Canada:
Institute Press.
Moos, R. H. (1993). Coping Responses Inventory: CRI-adult form professional manual. Odessa,
FL: Psychological Assessment Resources, Inc.
Neumann, D. A., Gallers, J., & Foy, D. W. (1989). The relationships between traumatic violence
and PTSD symptoms in rape victims. Unpublished master’s thesis, Fuller Theological Semi-
nary, Graduate School in Psychology, Pasadena, CA.
Nezu, A. M., & Carnevale, G. J. (1987). Interpersonal problem solving and coping reactions in
Vietnam veterans with posttraumatic stress disorder. Journal of Abnormal Psychology, 96,
155-157.
Parrot, A., & Bechofer, L. (1991). Acquaintance rape. New York: John Wiley.
Peterson, C., Semmel, A., von Baeyer, C., Abramson, L., Metalsky, G., & Seligman, M. (1982).
The Attributional Style Questionnaire. Cognitive Therapy and Research, 6, 287-299.
Raykov, T. (1994). Introduction to the structural equation modelling methodology in psychology
and the behavioral sciences. Lecture notes and tutorial exercises. Melbourne, Australia:
Department of Psychology, University of Melbourne.
Solomon, S. D. (1986). Mobilizing social support networks in times of disaster. In C. G. Figley
(Ed.), Trauma and its wake (pp. 43-52). New York: Brunner/Mazel.
Solomon, Z., Mikulincer, M., & Flum, H. (1988). Negative life events, coping responses, and
combat-related psychopathy: A prospective study. Journal of Abnormal Psychology, 97,
302-307.
Tennen, H., & Affleck, G. (1990). Blaming others for threatening events. Psychological Bulle-
tin, 107, 209-232.
Walker, L. E. (1984). Battered women’s syndrome. New York: Springer.
Wortman, C., & Lehman, D. (1985). Reactions to victims of life crisis: Support attempts that fail.
In I. G. Sarason & B. R. Sarason (Eds.), Social support: Theory, research, and applications
(pp. 463-489). Martinus Nijhoff: The Hague, the Netherlands.
Wyatt, G. E., Newcomb, M., & Notgrass, C. M. (1991). Internal and external mediators of
women’s rape experiences. In A. Wolbert Burgess (Ed.), Rape and sexual assault (pp. 32-43).
New York: Garland.

Dr. Suzanne L. Barker-Collo obtained her doctoral degree in clinical psychology from
Lakehead University. Her research interests include the evaluation of cognitive and
behavioral assessment and intervention strategies including their application to cogni-
tive rehabilitation and special populations (i.e., developmental disabilities and brain
injury). She holds a lectureship in neurorehabilitation in the Department of Psychology,
University of Auckland.

Dr. William T. Melnyk has 33 years experience in clinical psychology as a teacher,

researcher, administrator, and clinician. He has been a member of provincial, national,
and international committees including the Ontario Board of Examiners in Psychology
and the Association of State and Provincial Psychology Boards. He has taught in statis-
392 JOURNAL OF INTERPERSONAL VIOLENCE / April 2000

tics and design, psychometrics, abnormal psychology, cognitive behavior modification,

psychotherapy, and professional and legal issues. He has published and presented more
than 35 articles and led more than 50 workshops. He runs an extensive private practice.

Dr. Leslie McDonald-Miszczak obtained her doctoral degree in developmental psychol-

ogy at the University of Victoria. She currently holds an associate professorship in the
Department of Psychology at Western Washington University.